A Bayesian Approach for Detecting Mass-Extinction Events When Rates of Lineage Diversification Vary

Abstract

Abstract.-The paleontological record chronicles numerous episodes of mass extinction that 14 severely culled the Tree of Life. Biologists have long sought to assess the extent to which 15 these events may have impacted particular groups. We present a novel method for 16 detecting mass-extinction events from phylogenies estimated from molecular sequence data. 17We develop our approach in a Bayesian statistical framework, which enables us to harness 18 prior information on the frequency and magnitude of mass-extinction events. The approach 19 is based on an episodic stochastic-branching process model in which rates of speciation and 20 extinction are constant between rate-shift events. We model three types of events: (1) 21 instantaneous tree-wide shifts in speciation rate; (2) instantaneous tree-wide shifts in 22 extinction rate, and; (3) instantaneous tree-wide mass-extinction events. Each of the events 23 is described by a separate compound Poisson process (CPP) model, where the waiting 24 times between each event are exponentially distributed with event-specific rate parameters. 25The magnitude of each event is drawn from an event-type specific prior distribution. 26 Parameters of the model are then estimated using a reversible-jump Markov chain Monte 27Carlo (rjMCMC) algorithm. We demonstrate via simulation that this method has 28 substantial power to detect the number of mass-extinction events, provides unbiased 29 estimates of the timing of mass-extinction events, while exhibiting an appropriate (i.e., 30 below 5%) false discovery rate even in the case of background diversification rate variation. 31Finally, we provide an empirical application of this approach to conifers, which reveals that The copyright holder for this preprint (which was not . http://dx.doi.org/10.1101/020149 doi: bioRxiv preprint first posted online May. 31, 2015; The paleontological record documents numerous episodes of mass extinction that 40 severely culled the Tree of Life. As biologists, we often wish to assess whether these events 41 may have impacted particular groups. To this end, several statistical phylogenetic 42 approaches have been proposed to identify mass-extinction events from estimated 43 molecular phylogenies with divergence times, e.g., Nee et al. (1994) and Harvey et al. 44 (1994). These methods generally assume that the dated phylogeny is known without error, 45 and decompose the tree into a vector of waiting times between speciation events. Various 46 models of lineage diversification are then fit to these phylogenetic 'observations' using 47 maximum likelihood to estimate the speciation, b, and extinction, d, rates, and to identify 48 tree-wide shifts in diversification rate, including episodes of mass extinction. 60In order to distinguish more prosaic temporal variation in diversification rate from 61 bona fide mass-extinction events, we adopt a Bayesian statistical framework that enables us 62 to leverage prior information on the frequency and magnitude of mass-extinction events. 63Specifically, we d...

Increased diversification rates are coupled with higher rates of climate space exploration in Australian Acacia (Caesalpinioideae)

“…We then performed posterior predictive simulation using MCMC to assess the fit of the selected branching process model to the Acacia phylogeny. We explored the candidate branching‐process space including mass‐extinction events, using reversible jump MCMC implemented in C o MET (May et al , ). We used empirical hyper‐priors and ran the analysis for a maximum of 100 million iterations or a minimum effective sample size of 500, whichever came first.…”

Mass turnover and recovery dynamics of a diverse Australian continental radiation

“…To test for temporal variation in diversification rates, including Oligocene mass turnover, we used CoMET (May et al. ), as implemented in TESS (Höhna et al. ).…”

“…We applied the same Bayes factor significance thresholds as above and as have been used previously (May et al. ).…”

Insights on the evolutionary origin of Detarioideae, a clade of ecologically dominant tropical African trees

“…In order to calculate speciation and extinction rates, simulations were performed using the R package tess (Höhna, ; May et al ., ) and a tree with all outgroup taxa pruned. Empirical hyper‐priors were selected for a compound poisson process on mass extinction times (CoMET) analysis that was performed with 10 7 iterations.…”